Algorithm Research And System Implementation Of TCM Auxiliary Diagnosis And Treatment Based On Integrated Classification Model

Traditional Chinese medicine has a long history and has experienced thousands of years of historical accumulation and development.It is a treasure of Chinese traditional culture.The theoretical system of traditional Chinese medicine is a precious wealth that has been gradually formed by the people of all ethnic groups in China through the continuous accumulation of experience in the struggle against diseases and has been passed down from generation to generation.At present,there is an obvious imbalance between supply and demand between the development scale,service quality and the health needs of the people of traditional Chinese medicine,and it is also facing the dilemma of inheritance.At present,there is an obvious imbalance between supply and demand between the development scale of traditional Chinese medicine,service quality and the health needs of the people,and it is also facing the dilemma of traditional Chinese medicine inheritance.This thesis uses machine learning technology to build a TCM integrated classification model by learning and exploring the diagnosis and treatment knowledge and rules contained in the TCM classic "Treatise on Febrile Diseases".Based on this model,a TCM auxiliary diagnosis and treatment system is implemented to help doctors quickly and accurately diagnosis and treatment decisions.The main work of this thesis is as follows:(1)Construct an integrated classification model of traditional Chinese medicine.Using the six-channel syndrome differentiation thinking in "Treatise on Febrile Diseases",the diagnosis and treatment data of "Treatise on Febrile Diseases" were analyzed,and important symptom features were screened out through feature selection based on chisquare test,feature selection based on random forest and recursive feature elimination.The Voting algorithm integrates the base classifiers of KNN,SVM,and CART decision trees to realize the integrated classification model.Experiments are carried out on the public "Treatise on Febrile Diseases" diagnosis and treatment dataset.The indicators of the integrated classification model are better than those of the base classifier,and the accuracy rate reaches 67.88%.(2)A TCM tongue color recognition algorithm is proposed to assist the integrated classification model of TCM.Considering that the symptoms in the diagnosis and treatment data set of "Treatise on Febrile Diseases" will include tongue color categories,this thesis uses the maximum inter-class variance segmentation algorithm to segment the tongue body and implements the tongue color recognition algorithm through the neural network built by Le Net-5.In the TCM tongue image data Experiments were carried out on the dataset and public datasets,and the accuracy rate reached 94.57%.(3)Design and implement TCM auxiliary diagnosis and treatment system.Based on the integrated classification model of traditional Chinese medicine and the tongue color recognition algorithm of traditional Chinese medicine,the functions of syndrome discrimination,tongue color recognition and prescription recommendation are jointly realized.According to the requirements of software engineering,the whole process of demand analysis,system design,system development and system testing is gradually completed,and finally realized A TCM auxiliary diagnosis and treatment system.